Crowdsourcing Semantic Data

Per definition semantic technologies aim to reach a high level of automation with respect to the management and use of an open and decentralized data space bringing together data sources published by different parties, of varying quality and using heterogeneous schemes and vocabularies. This is achieved by using established Web technologies known for their robustness and scalability, paired with a uniform data model and links between datasets facilitating exploration and integration, as well as explicit conceptual models known as ontologies, which describe the structure and properties of the data in various domains. Nevertheless, while the core technological building blocks of this approach are maturing, and large amounts of useful data are continuously made available on the Web, the actual experience in developing semantic applications reveals that many aspects of semantic content management remain, for principled or technical reasons, heavily reliant on human intervention. In this session we will have a look at a prominent selection of such semantic content management aspects and analyze the challenges and opportunities brought in by different crowdsourcing approaches such as games-with-a-purpose and microtask platforms. We will introduce a series of developments of semantic technologies and tools which leverage the power of human computation for tasks such as ontology development, ontology alignment, semantic annotation, and SPARQL query processing.